PHARMACODYNAMICS OF SALTS AND DRUGS 113 



titative comparisons. It will be noticed that nickel is more poisonous 

 than cobalt for Pisum sativum, and far more poisonous than cobalt 

 for Zea mais. Both of these roots appear less sensitive to acids 

 than to the metals, the hydrogen ion being less toxic than nickel. 

 With Lupinus, while the general order is the same as that observed 

 elsewhere, cadmium is here more poisonous than copper just 

 the exception noted for Fundulus. 



The results of Clark and Stevens upon mold spores are also for the 

 purposes of quantitative comparisons unsatisfactory. These spores 

 appear to be surrounded by such membranes, or to have so great 

 a resistance, as to make the interpretation of results very doubtful, 

 although there can be no doubt that the general trend of the results 

 is the same as that already noted. 



True and Kahlenberg's results (Table i) are more satisfactory, 

 but, owing to the fact that these authors did not study many of the 

 metals, and also did not pretend to fix the fatal point with accuracy, 

 their results are not satisfactory for quantitative treatment. True's 

 results upon the toxicity of the salts of the various acids are unfortu- 

 nately unavailable for our purposes, owing to the uncertainty of 

 the ionic potential of these anions. 



/ ) The solubility of globulin in salt solutions. In a previous 

 paper 1 I showed that the solubility of sodium albuminate (egg albu- 

 min in alkaline solution) in different salt solutions was determined 

 by the tension-coefficient of the salt. By the tension-coefficient 

 was meant the difference between the solution tensions of the ions 

 divided by their sum. The numerator of this fraction should be 

 the ionic potentials instead of the solution tensions. I showed that 

 in an alkaline solution the solubility was greater in sodium iodide 

 than in the bromide or chloride, and that when the ionic potential 

 of the cation surpassed a certain figure the salts precipitated the 

 albumin. Table II 3 shows the relationship (qualitative) between 

 the ionic potentials of salts and their power of dissolving or pre- 

 cipitating such albumin, both the unboiled and the boiled. It 

 will be seen by an inspection of the table that the ionic potential 

 arranges the salts in the order of their action on the albumin. 



Osborne and Harris 2 have estimated quantitatively the solvent 



'MATHEWS, Amer. Jour. Physiol., 1905, 14, p. 204. 

 ' OSBORNE AND HARRIS, ibid., 1905, p. 151. 

 J MATHEWS, loc. cit., p. an. 



